Extended Data Fig. 8: Experimental links between structure and function of the precatalytic complex.

a) The structural features of the scaffolding involve close contacts between the two binding arms, non-base pairing at position +1, and metal-ion binding in region I. b) The clear absence of respective peaks in the NOE pattern confirming flip out of dG₊₁. Unlike to all nucleotides with expected in-register stacking, no cross correlations of dG₊₁ to dG₊₂ is detected (which should appear along the dashed blue line). Stacking pattern for dA₊₆ – dT₊₉ is shown as positive control (green arrows). c) Effects of mutations at positions +1 and −1 on Dz^5C activity confirming that position +1, unlike position −1, does not form an essential Watson-Crick base pair. d) Example of NMR data (extracts of NOESY spectra) confirming close spatial proximity of dG₋₅ and dG₊₂ after Mg²⁺ binding. e) Atom-specific PRE rates obtained from Mn²⁺ titration for DNA^c–RNA (top) compared to respective nucleotides in Dz^5C–RNA^2ʹF (bottom). The data demonstrate that the presence of the catalytic loop dramatically changes the M²⁺-binding behaviour of the arms from a rather diffuse pattern (top) to the defined binding region that forms the basis of the scaffolding step (bottom). f) Top view on the precatalytic structure focusing on the cleavage-site surroundings. g) NOE-buildup rates strongly indicate syn-conformation of dA₋₁. The schematic model shown on top visualizes the considered inter-proton interactions with either fixed distances (H1ʹ-H2ʹ, purple, and H1ʹ-H2 ″, red) or distances strongly depending on the χ-angle (H1ʹ to indicated base proton, blue). h) NOE pattern indicative of the Mg²⁺-induced flip out of T4. Note the same behaviour is observed for all ten resolved inter-nucleotide correlations of T4 to its neighbours. i) Changes of cleavage activity by mutations in the 5ʹ side of the catalytic loop (metal ion-binding site II). Mutations at position 5 serve as a reference; all other mutations are variants of Dz^5C. Data are presented as mean values +/- SD of triplicate experiments. j) Comparison of the cleavage activity of the variants Dz^5A (black), Dz^5C (blue) and Dz^5G (red) in the presence of 3 mM Mg²⁺ (data for Dz^5A and Dz^5C are identical to respective conditions shown in Fig. 1b). k) Simplified schematic model highlighting the central features of the 3D structure. l-n) cw EPR characterization of Mn²⁺ binding. l) Double integrated EPR signal of Mn²⁺ in the absence (blue) and presence of 40 μM Dz^5C·RNA^2ʹF complexes (black). Full range and zoom into lower concentrations are shown. m,n) Fit of experimentally determined binding behaviour (black data points) for either full Mn²⁺ concentration range (left panels) or only the higher affinity binding sites (right panels). Data were fitted to different binding models (red curves, see Methods and Supplementary Discussion for more details on the applied data fitting). n) Three cooperative binding sites were assumed, resulting in high- and low-affinity binding sites and overall the best fit to the obtained data (left). In the right panel, a model with only one cooperative binding site was assumed, and only the shown data range was considered. The shown fit captures the key features of the binding behaviour with minimal number of parameters.